Unverified50% confidenceSolutionExact time
After catching an exception, the error information should be returned to the large model verbatim rather than silently discarded, enabling the model to self-reflect, adjust its format, and retry, thereby improving agent robustness.
1
Sources
50%
Confidence
Long-term
Relevance
9/26/2026
First Seen
Sources
Related Entities
Related Claims
Unverified工具调用使攻击面从让模型说错话升级为让模型做错事，危害从信息误导上升为实质性的数据访问和系统篡改76% similarUnverifiedThe results of human review should flow back to continuously calibrate the model and costs; otherwise the system will fail as data drifts.74% similarUnverifiedWhen a model fails to correctly invoke retrieval tools or determines that a question does not require a search, it falls back to answering from training data, thereby exposing the knowledge cutoff issue.73% similarUnverifiedAn agent-style iterative error-correction loop that feeds error messages back to the model for self-correction when query execution fails can significantly improve the accuracy of the final results.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/952344API
curl https://kongchang.com/api/v1/knowledge/claims/952344MCP
get_claim(id=952344)